Effect of Allopurinol on Postasphyxial Free Radical Formation, Cerebral Hemodynamics, and Electrical Brain Activity
Objective. Free radical-induced postasphyxial reperfusion injury has been recognized as an important cause of brain tissue damage. We investigated the effect of high-dose allopurinol (ALLO; 40 mg/kg), a xanthine-oxidase inhibitor and free radical scavenger, on free radical status in severely asphyxiated newborns and on postasphyxial cerebral perfusion and electrical brain activity.
Methods. Free radical status was assessed by serial plasma determination of nonprotein-bound iron (μm), antioxidative capacity, and malondialdehyde (MDA; μm). Cerebral perfusion was investigated by monitoring changes in cerebral blood volume (ΔCBV; mL/100 g brain tissue) with near infrared spectroscopy; electrocortical brain activity (ECBA) was assessed in microvolts by cerebral function monitor. Eleven infants received 40 mg/kg ALLO intravenously, and 11 infants served as controls (CONT). Plasma nonprotein-bound iron, antioxidative capacity, and MDA were measured before 4 hours, between 16 and 20 hours, and at the second and third days of age. Changes in CBV and ECBA were monitored between 4 and 8, 16 and 20, 58 and 62, and 104 and 110 hours of age.
Results. Six CONT and two ALLO infants died after neurologic deterioration. No toxic side effects of ALLO were detected. Nonprotein-bound iron (mean ± SEM) in the CONT group showed an initial rise (18.7 ± 4.6 μm to 21.3 ± 3.4 μm) but dropped to 7.4 ± 3.5 μm at day 3; in the ALLO group it dropped from 15.5 ± 4.6 μm to 0 μm at day 3. Uric acid was significantly lower in ALLO-treated infants from 16 hours of life on. MDA remained stable in the ALLO group, but increased in the CONT group at 8 to 16 hours versus <4 hours (mean ± SEM; 0.83 ± 0.31 μm vs 0.50 ± 0.14 μm). During 4 to 8 hours, ΔCBV–CONT showed a larger drop than ΔCBV–ALLO from baseline. During the subsequent registrations CBV remained stable in both groups. ECBA–CONT decreased, but ECBA–ALLO remained stable during 4 to 8 hours of age. Neonates who died had the largest drops in CBV and ECBA.
Conclusion. This study suggests a beneficial effect of ALLO treatment on free radical formation, CBV, and electrical brain activity, without toxic side effects.
Hypoxia and ischemia during severe birth asphyxia result in inadequate substrate supply to meet the metabolic demands leading to destruction of susceptible brain cells. Although reperfusion is necessary to prevent additional damage to neuronal tissue, the return of oxygenated blood to previously ischemic brain tissue has been increasingly recognized as an important mechanism for substantial additional brain injury and includes both damage to the microcirculation and parenchyma of the brain. This injury is in large part mediated by free radical formation on post hypoxic–ischemic reoxygenation.1-3 Damaging amounts of superoxide, hydrogen peroxide, and the very toxic hydroxyl radical are generated by free fatty acids and prostaglandin metabolism.3,4 Moreover, metabolization of hypoxanthine, accumulated in the brain during hypoxia and ischemia, and by xanthine oxidase, which is concentrated precariously within the endothelium of the cerebral microvessels, leads to an additional production of superoxide and hydroxyl radicals.5
Several studies in adult and newborn animals report that allopurinol (ALLO) and its active metabolite oxypurinol offer protection against reperfusion-induced brain injury by reducing free radical formation.6,7 Although most studies pretreated the animals with ALLO, as a rescue therapy ALLO also showed reduction of hypoxia-induced brain injury.8 The protective action of this drug has been explained primarily by its xanthine oxidase-inhibiting property, but at high concentrations, it also scavenges free radicals such as hydroxyl and chelates transition metals like nonprotein-bound iron.9,10 Earlier studies in preterm and term infants using 20 mg/kg per day showed no adverse effects of ALLO.11-13 The main aim of the present study therefore was to investigate the hypothesis that early neonatal ALLO treatment with higher dosages (40 mg/kg in 12 hours) decreases free radical formation and lipid peroxidation after severe birth asphyxia in the human newborn without producing toxic side effects.
Hemodynamically and functionally, reperfusion injury of the brain after severe birth asphyxia has been characterized by cerebral hypoperfusion and a reduced electrical activity of the brain.14-16 The second aim of the present study therefore was to investigate the hypothesis that ALLO treatment preserves brain perfusion and electrical brain activity. Changes in brain perfusion were assessed with near-infrared spectroscopy (NIRS).17 Electrocortical brain activity (ECBA) was determined with cerebral function monitoring.18
PATIENTS AND METHODS
Newborn infants admitted consecutively to the neonatal intensive care unit between February 1995 and March 1996, who suffered from severe birth asphyxia, were enrolled in the study. They met the following criteria: 1) gestational age ≥35 weeks, determined by maternal dates and/or Ballard score;19 2) admission into our neonatal unit within 2 hours after birth; 3) suffering from severe birth asphyxia defined as fetal distress (abnormal heart rate pattern, meconium-stained amniotic fluid, and a cord or first pH of <7.00) requiring immediate neonatal ventilation with mask or endotracheal tube for >2 minutes; and 4) no major congenital anomalies. The infants received either 40 mg/kg ALLO intravenously or served as controls (CONT). Randomization was determined by opening numbered sealed envelopes containing cards indicating whether ALLO treatment would be performed. Because we used high dosages of ALLO, we decided not to blind the drug to monitor possible side effects. The side effects of ALLO we focused on were abnormal changes in total and differential white blood counts and in the platelet count. Furthermore, we looked for cutaneous reactions possibly caused by ALLO such as pruritic, erythematous, or maculopapular eruptions. In this respect, special attention was given infants receiving both antibiotics (amoxicillin) and ALLO. The study was approved by the scientific board of the Department of Pediatrics and the ethical committee of the University Hospital of Leiden. Informed parental consent was obtained in all cases after oral and written explanation of the aim of the study.
Measurement of Prooxidant Activity, Antioxidant Capacity, and Lipid Peroxidation
Blood was collected into heparinized glass tubes and centrifuged immediately (750 × g, 10 minutes); the plasma was stored under argon at −70°C until analysis. Plasma samples that showed pink discoloration (hemolysis) were excluded from the study. Nonprotein-bound iron, a prooxidant, was measured using the bleomycin assay.20 Using this assay, the absence of nonprotein-bound iron, ie, the presence of iron-binding capacity, can be measured as well as the presence of nonprotein-bound iron, ie, the lack of iron-binding capacity attributable to saturation or dysfunction of transferrin. If nonprotein-bound iron is present, the lower detection limit is 0.6 μm. The glass tubes used to collect the blood did not contain detectable amounts of iron. The intra- and interassay coefficients of variation of the bleomycin assay were 6.6% and 7.4%, respectively.
High-performance liquid chromatography techniques were used to determine ascorbic acid (AA) and its oxidation product dehydroascorbic acid (DHA),21 and uric acid and its oxidation product allantoin.22 We also measured the total peroxyl radical trapping antioxidant capacity of plasma (TRAP). This assay measures the synergistic action of the chain-breaking antioxidants in plasma to prevent lipid peroxidation.23
The lipid peroxidation product malondialdehyde was measured using high performance liquid chromatography.24 All measurements were performed within 6 weeks after collection of the sample.
Assessment of Cerebral Hemodynamics by NIRS
The head of the neonate is relatively transparent to NIRS. Hemoglobin (Hb) and cytochrome oxidase (the terminal member of the mitochondrial respiratory chain) are natural chromophores and have both an oxygenation-dependent absorption in this wavelength region. By selection of appropriate wavelength, an algorithm has been developed to convert absorption changes into changes of oxygenated Hb (ΔHbO2), deoxygenated Hb (ΔHbR) and total Hb (ΔHbtot) (= ΔHbO2 + ΔHbR).17The NIRS instrument used (Critikon 2000, Johnson & Johnson Medical Ltd, Norderstedt, Germany) consisted of four laser sources with wavelengths of 776.5, 819.0, 871.4, and 908.7 nm. The maximal power per laser was 10 W peak. The average optical power into tissue was <5 mW. The neonatal sensor that contains a photo diode and light detector separated 3.5 cm from each other was placed over the occipital brain region in the left parasagittal plane and attached firmly to the skull with stretch bandages. The calculated concentration changes of ΔHbO2, ΔHbR, and ΔHbtot were expressed in micromoles. The algorithm incorporates wavelength-dependent differential pathlength factor data derived from time of flight studies.25
Assuming a stable hematocrit, changes in ΔHbtot will reflect changes in cerebral blood volume (ΔCBV). Relative changes in ΔCBV were calculated using the following equation; ΔCBV = ΔHbtot × 0.89/[Hb], where [Hb] is the large vessel Hb concentration in grams per deciliter.26 Changes in ΔCBV thus are relative changes from the baseline value and expressed in milliliters per 100 g of brain tissue. Earlier studies showed a good relationship with changes in actual brain blood flow, determined with the 133xenon clearance method.17Therefore, we considered that ΔCBV assessed qualitatively changes in cerebral blood flow, if changes in ΔCBV were caused predominantly by changes in ΔHbO2. However, it is important to realize that monitoring of ΔCBV over such long periods (∼4 hours) may negatively influence its relation with brain perfusion.
Determination of ECBA
Changes in ECBA were monitored using a filtered and selectively amplified one channel cerebral function monitor (Lectromed, Oxford Instruments, Oxford, UK), as described by Prior.27 The cerebral function monitor has a special filter that sharply attenuates frequencies <2 and >15 Hz, giving an amplitude-integrated recording that contains the main electroencephalographic frequencies, but with little disturbance from artifacts. The electroencephalographic signal was obtained from a pair of silver-chloride disk electrodes placed with electrode cream at the P3 and P4 position of the 10 to 20 International System, ie, in the left and right parietal region. The ECBA was recorded on a semilogarithmic scale (0 to 100 μV). The paper speed was 2 mm per minute. Simultaneously with the amplitude curve, an impedance curve recorded the reliability of the signal by a reference electrode positioned anteriorly on the scalp and showed artifacts from movement or loose electrodes.
Clinical Data of the Studied Infants
Obstetric and intrapartum data were collected from hospital records. Neonatal data were collected prospectively. Clinical management decisions were made by the attending neonatologist, and no attempts were made to influence the clinical care.
The Sarnat classification28 was used for defining three clinical stages of hypoxic–ischemic injury of the brain due to birth asphyxia. Hypoxic–ischemic encephalopathy (HIE) stage I included hyperalertness, hyperreflexia, and tachycardia; HIE stage II included lethargy, hyporeflexia, bradycardia, hypotonia, weak suck and Moro reflexes, and convulsions; HIE stage III included stupor, profound hypotonia, hypothermia, and absent suck and Moro reflexes.
Samples for blood gas analysis and hematocrit were obtained from an arterial line (umbilical, radial, or posterior tibial artery) or from arterialized capillary blood samples, at least once every 2 to 3 hours during the first 12 hours of age and every 6 hours thereafter. When assisted ventilation was necessary, time-cycled pressure-limited infant ventilators (Bourns BP 200, Bear Medical Systems Inc, Riverside, CA, or Infantstar, Infrasonics Inc, San Diego, CA) were used. Arterial pressure was determined with an indwelling arterial catheter or with an oscillometric method (Dynamap, Criterion, Tampa, FL).
Study Design, Data Collection, and Analysis
The infants treated received 40 mg/kg of ALLO (apurin, pH of solution: 10 to 11) by slow intravenous infusion over a 20 minute period; the first dose of 20 mg/kg had to be given before 4 hours of age, and the second dose 12 hours later. The study was not blinded for safety reasons (quick awareness of possible adverse reactions of ALLO). The CONT infants did not receive a placebo. The infants were connected to the NIRS apparatus and cerebral function monitor between 4 and 8, 16 and 20, 58 and 62, and 104 and 110 hours of age. We aimed to obtain registrations with a duration of 4 hours. Relative to the value at the onset of the recording, ΔHbO2, ΔHbR, Hbtot, (ΔCBV), and changes in ECBA were determined simultaneously every second and stored in a personal computer for off-line analysis. The redox status of the blood (prooxidant activity, ie, nonprotein-bound iron; antioxidant parameters; and lipid peroxidation), hematocrit, and blood count were assessed twice during the first 24 hours of life (before the age of 4 hours and between 16 and 20 hours of age), and once at days 2 and 3 in venous or arterial blood samples. Liver enzymes (serum glutamic oxaloacetic transaminase, serum glutamic pyruvic transaminase, and lactic acid dehydrogenase); renal function (creatinine, serum urea nitrogen); and CPK, including CK-BB fraction percent of total in case CPK was elevated) were determined at day 2. This time point was chosen because earlier determinations of some of these chemical parameters may reflect maternal rather than neonatal values. Blood gasses were determined at regular intervals. In infants receiving ALLO, plasma concentrations of this drug and its active metabolite oxypurinol were determined 2 hours after the first dose, just before the second dose, and 2 hours after the second dose. The HIE staging with the Sarnat score was performed by the attending neonatologist before 4 hours of age and before ALLO administration in those infants receiving this drug, and then daily or more frequently if an alteration in the patient's neurologic condition occurred. In those infants who survived, a routine neurologic examination was performed at discharge by the attending neonatologist, who was unaware of the study results.
Differences between groups regarding clinical and laboratory data were assessed by Student's t test, Wilcoxon's rank sum test, or χ2 test, where appropriate.
To investigate differences between prooxidant activity, antioxidant parameters, and lipid peroxidation values of the two groups at each postnatal time point, one-way factorial analysis of variance was performed. When a significant difference was found, analysis of variance was followed by the Scheffe's procedure. To investigate differences within a group as a function of postnatal age for these variables, multiple linear regression analysis was performed.
For statistical evaluation of the patterns of ΔCBV and ECBA and for their representation in the figures, the second-to-second values recorded for each patient studied were averaged every 10 minutes. To investigate the patterns of ΔCBV and ECBA as a function of postnatal age and whether ALLO had a significant effect on this relationship, we used a multiple linear regression model. In all four subsequent registration periods, we used the same multiple linear regression model with dummy variables with the following regression equation:Equation 1where ΔCBV (1) or ECBA (2) were the dependent variables (Y1 or Y2) andao (ao1 orao2) their means over all the runs. Registration time T is the first independent variable, and its coefficient aT defines the slope of the relationship between the dependent variable (ΔCBV or ECBA) and postnatal time T. The second independent variable is the (dummy) drug variable A (ALLO), representing the CONT group (value = −1) or the ALLO-treated group (value = +1). The coefficient aA indicates the independent effect of ALLO treatment on the dependent variable (ΔCBV or ECBA), thus affecting the intercept of the relationship between the dependent variable and the postnatal time T. The third independent variable is an interaction variable, T * A, representing the interactive effect of ALLO and postnatal time on the dependent variable (ΔCBV, Δcytochrome aa3 or ECBA). Therefore the coefficient aT * A indicates the effect of ALLO on the slope of the dependent variable–postnatal time relationship. Finally, to correct for interpatient variability, 21 dummy variables (P1–P21) were introduced for the 22 infants included in this analysis. A more detailed explanation of this type of statistical analysis has been given elsewhere.29
P values < .05 were considered statistically significant.
The total number of infants studied was 22; 11 received ALLO (median postnatal age at the first dose was 170 minutes, ranging from 68 to 210 minutes), and 11 infants served as controls. Table1 shows relevant clinical data (including the Sarnat score before 4 hours of age and before ALLO treatment), maximal Sarnat score during the study period, and adverse short-term outcome. Maximal Sarnat scores tended to be lower in the ALLO group. Ten of 11 infants in the CONT group and 8 of 11 infants in the ALLO-treated group needed artificial ventilation beyond the resuscitation period. Seven infants of the CONT group and 5 infants of the ALLO group needed anticonvulsive therapy. Medication was started within the first 6 hours of life in 11 of the 12 infants. One infant in the ALLO group had seizures on day 4. Phenobarbital was the drug of choice (starting dose up to 30 mg/kg; maintenance, 3 to 5 mg/kg per 24 hours). In case seizures continued despite adequate phenobarbital medication, phenytoin (starting dose, 15 mg/kg; maintenance, 3 to 5 mg/kg per 24 hours) and clonazepam (starting dose, 0.1 mg/kg; maintenance, 0.1 to 0.5 mg/kg per 24 hours as continuous infusion) were added. Four and 3 infants in the CONT and ALLO groups, respectively, needed positive inotropic support to maintain blood pressures within normal limits (>40 mm Hg) in the first 12 hours of life. Table2 shows mean blood pressure, arterial Pco2, O2saturation, and pH (mean ± 1 SD) of both groups during the subsequent registration periods of CBV and ECBA. Blood pressures, arterial Pco2, O2 saturation and pH, and hematocrit (data not shown) were stable, not different between groups, and within normal limits during the subsequent registration periods.
Liver enzymes, renal function tests, and CPK were elevated equally in both groups on the second day of life (Table 3). Liver enzymes and renal function tests were determined on days 3 and 5 in those infants of both groups who still were admitted; liver enzymes and renal function tests recovered similarly in all infants to normal values. Urine output was low in both groups during the first 24 hours of life (CONT [mean ± 1 SD], 1.0 ± 0.7; ALLO, 0.8 ± 0.6 mL/kg per hour), but recovered in the surviving infants during the subsequent days (CONT [mean ± 1 SD], 3.2 ± 0.6; ALLO, 3.6 ± 1.1 mL/kg per hour).
Figure 1 shows the subsequent individual ALLO concentrations in plasma; mean values ± 1 SD and ranges were 16.6 ± 3.0 (11.1 to 20.0), 7.4 ± 5.2 (3.0 to 18.7), and 29.9 ± 9.9 (19.1 to 51.5) μg/mL at 2 hours after the first dose, just before the second dose, and 2 hours after the second dose, respectively. After the second dose of ALLO, nearly all infants had very high plasma levels of this drug (upper limit of therapeutic concentration in plasma, 13.6 μg/mL). However, none of the infants treated with ALLO showed any adverse reactions. Total white blood counts, differential white blood counts, and platelet counts at admission (before ALLO treatment), at 16 to 20, 48, and 72 hours were within normal ranges in almost all patients and did not differ between groups. No patient showed skin manifestations that could be linked to the use of ALLO. Moreover, although all infants showed initially elevated liver enzymes, no differences were found between groups, making it unlikely that ALLO increased further the transaminase activities. Oxypurinol levels (data not shown) were mostly very low or not detectable.
Five CONT infants died after deterioration of their neurologic condition (two infants died during the first day of life), one died before 12 hours of age because of hypoxia-induced cardiac failure. In the ALLO-treated infants only two died after neurologic deterioration (one on the first day). Infants with adverse outcome had the highest maximal Sarnat scores (2 or 3). At discharge, all but one of the surviving infants had a normal neurologic examination; one CONT infant was slightly hypertonic but otherwise neurologically normal at discharge. However, significance between the groups with respect to short-term outcome was not reached. Three surviving CONT and two surviving ALLO infants were on anticonvulsant medication when discharged.
Prooxidant Activity, Antioxidant Parameters, and Lipid Peroxidation
Figure 2 shows the values (mean ± SEM) of nonprotein-bound iron, uric acid, TRAP, and malondialdehyde as a function of postnatal age.
Nonprotein-bound iron plasma concentration was not different between groups before ALLO treatment. In the CONT group nonprotein-bound iron showed an initially small increase, with a sustained nonsignificant decrease afterward (18.7 ± 4.6, 21.3 ± 3.4, and 7.4 ± 3.5 μm before 4 hours, at 16 to 20 hours, and during the third day, respectively). Nonprotein-bound iron in the ALLO group, however, decreased to virtually 0 with increasing postnatal age (15.5 ± 4.6 and 1.4 ± 0.9 μm before 4 hours and during the third day, respectively; P < .0001). At the second and third days of life, nonprotein-bound iron was significantly lower in the ALLO group compared with the CONT group.
A stable pattern of AA/DHA ratio (data not shown) as a function of postnatal age was found in CONT as well as in ALLO infants, and no consistent differences were seen between groups. Allantoin in the CONT group was somewhat higher at 16 to 20 hours of age and on day 2 compared with the ALLO group. However, unlike in the ALLO group, it decreased significantly on day 3 (CONT, 17.5 ± 4.2 at <4 hours, 17.7 ± 4.6 on day 2, and 5.2 ± .9 μm on day 3; P < .05; ALLO, 15.7 ± 4.4 at <4 hours, 13.7 ± 9.5 on day 2, and 9.6 ± 4.2 μm on day 3; not significant). No differences between groups were detected. In both groups, uric acid decreased significantly over time (717 ± 81 to 327 ± 113 and 557 ± 23 to 159 ± 15 μm at <4 hours and day 3 of age, respectively), but was significantly lower during all time points in the ALLO group compared with the CONT group (P < .05). The allantoin/uric acid ratio therefore was rather stable in the CONT group (2.33 ± 1.31 at <4 hours of age vs 2.98 ± 1.40% on day 2), but showed a nonsignificant increase in the ALLO group over time (2.91 ± 2.16 at <4 hours of age vs 7.37 ± 6.68% on day 2). The TRAP, an indicator of the overall chain-breaking antioxidative capacity of the plasma, was initially stable in both groups, but showed a sustained nonsignificant decrease from 16 to 20 hours of life to the third day (1364 ± 127 to 855 ± 221 μm for the CONT group; 1007 ± 106 to 681 ± 67 μm for the ALLO group). TRAP values in the ALLO group were significantly lower compared with the CONT group during day 2 (P < .05).
Malondialdehyde showed a significant increase (0.50 ± 0.14 to 0.83 ± 0.31 μm) at 16 to 20 hours postnatally compared with 4- to 8-hour values in the CONT group, but decreased and remained stable afterward. The ALLO group values remained stable over postnatal time (0.51 ± 0.15 to 0.61 ± 0.8).
Patterns of ΔCBV During Subsequent Registration Periods
Because of technical problems, we have no complete NIRS registrations from two patients in the CONT group and one in the ALLO group.
Figure 3A shows the changes in CBV during 4 to 8 hours of age. Both groups showed a sustained drop over time in the majority of infants, although this trend was more pronounced in the CONT group. Especially those CONT infants with adverse outcome, indicated by an arrowhead in Fig 3, showed the greatest drop. Multiple linear regression analysis indeed showed a relationship between postnatal age and ΔCBV during the 4- to 8-hour registration period for both groups, a significant decrease compared with baseline for CBV of 0.003 and 0.001 mL/100 g per minute for the CONT group (P < .0001) and ALLO group (P < .01), respectively. As indicated by these numbers, ALLO did indeed affect significantly the intercept (P < .0001) and the slope (P < .01) of the ΔCBV postnatal time relationship during this 4- to 8-hour registration period, indicating a greater and steeper decrease of CBV over time in the CONT group compared with the ALLO group. It appeared additionally that infants with the highest Sarnat score had the greatest decrease in CBV. During the subsequent registration periods, ie, between 16 and 20, 58 and 62, and 106 and 110 hours of age, patterns of ΔCBV were stable in both groups, and no significant change of ΔCBV occurred in relation with the registration time.
ECBA During Subsequent Registration Periods
Because of technical problems or incorrect calibration, no reliable ECBA tracings were available in three CONT infants and four ALLO infants during one (n = 2), two (n = 2), three (n = 2), or all (n = 1) registration periods.
Figure 3B shows the ECBA patterns of both groups during the 4- to 8-hour registration period. In general, it is obvious from the tracings that the patients with adverse outcome show the lowest ECBA values compared with those infants with a favorable short-term outcome. Because only two of eight infants with an adverse outcome originated from the ALLO group, it is possible that the ALLO-treated infants had significantly higher ECBA values at all time points during the 4- to 8-hour registration period (P < .0001). In contrast, the lower ECBA in the CONT group seems primarily to be related to the greater number of infants with adverse outcome. Here again, infants with the highest Sarnat scores had the lowest ECBA values. ECBA tracings during the subsequent periods depict similar tracings (data not shown).
The small number of infants in the present study may have influenced negatively the homogeneity of the groups studied with respect to the severity of the birth asphyxia. There are, however, several reasons that suggest strongly that both groups had a comparable amount of hypoxic–ischemic stress and early reperfusion injury; umbilical artery pH values were comparable between groups; the Sarnat score before 4 hours of age and before ALLO treatment was similar between groups; and the initial redox status of the plasma of both groups, considered a rather sensitive measure of hypoxic–ischemic stress, showed an equal amount of abnormally high prooxidant activity (nonprotein-bound iron), comparable antioxidant capacities, and similar malondialdehyde concentrations. It is important to state that the concentrations of the plasma nonprotein-bound iron detected were abnormally high. Although earlier publications found detectable concentrations of nonprotein-bound iron in plasma in a minority of healthy term infants, these concentrations were much lower.23
With the above-mentioned considerations in mind, it is noteworthy that the ALLO-treated infants had significantly less production of nonprotein-bound iron, an important prooxidant, during the study period and showed no significant rise in malondialdehyde at 16 to 20 hours postnatally, contrary to those infants not treated with ALLO. With respect to a possible positive effect of ALLO on free radical production and, hence, on postasphyxial reperfusion injury of the brain, literature indeed supports the neuroprotective action of this drug. ALLO inhibits the enzyme xanthine-oxidase that converts hypoxanthine, produced in large amounts during the actual hypoxic–ischemic insult, in the postasphyxial reoxygenation phase to uric acid, thus producing large amounts of the free radical superoxide. Patt et al30 showed that inhibition of xanthine oxidase in a gerbil model of cerebral ischemia reduced free radical production and brain edema. Palmer et al7 reported that hypoxic–ischemic brain damage in immature rats could be prevented by pretreatment with high-dose ALLO. The same group found that ALLO also was protective as rescue treatment, administered up to 4 hours after the hypoxic insult,31,32 although there was a dose dependency.8
High plasma levels probably are required for the neuroprotective effects of ALLO. It has been suggested that at higher blood levels, ALLO and its metabolite oxypurinol can act as scavengers of the very toxic hydroxyl radical and as transition metal chelators.9,10 The last probability is supported by the present study when considering the nonprotein-bound iron data. There were significantly lower concentrations of this prooxidant in plasma of ALLO-treated infants, which gives rise to a substantial production of the very toxic hydroxyl radical via the iron catalyzed Fenton or Haber–Weiss reaction.33
It indeed has been suggested that ALLO may act primarily as a free radical scavenger and exerts its activity by inhibition of xanthine oxidase activity much less, which shows only a relatively low activity in most human organs.9,34 An additional reason to assume that the action of ALLO in this particular study focuses on free radical scavenging and less on inhibition of xanthine oxidase is the fact that the latter action is likely to occur early, within the first hours on reperfusion and reoxygenation. The free radical scavenging effect of ALLO, however, is more sustained. Because our study used ALLO as a postevent intervention (median postnatal age at first dose, 170 minutes), it is unlikely that xanthine oxidase inhibition by ALLO played a substantial role. Saugstad34 stresses further that ALLO drastically reduces uric acid concentration. Uric acid is one of the most important antioxidants, thus, this may reduce antioxidative capacity in the birth-asphyxiated newborn. Indeed, our data revealed significantly lower uric acid concentrations in the plasma after ALLO treatment, and it is probably also the cause of the lower TRAP values during the second day of life, indicating less peroxyl radical- trapping capacity. Thus, the advantages of ALLO therapy may be partially diminished by the lowering of the chain-breaking antioxidant capacity of the plasma. Measurement of plasma malondialdehyde and the powerful plasma antioxidants AA and uric acid, their oxidation products, and resulting respective ratios did not provide an answer to this important question. There were no differences in either plasma malondialdehyde or AA/DHA ratios. Furthermore, the allantoin level did not differ between groups, and therefore the higher allantoin/uric acid ratio cannot be interpreted as increased oxidative stress. It has been pointed out recently that this ratio may be of limited value as a marker of oxidative stress in the newborn, because therapy, ie, ALLO and changes in renal tubular function, can markedly effect the uric acid concentration.35
Although we used rather high dosages of ALLO in the present study, no adverse effects were detected. As presented in Fig 1, almost all infants showed ALLO concentrations in plasma >13.6 μg/mL, especially after the second dose. Above this concentration, toxic side effects of ALLO on the skin, white blood count abnormalities, and elevated liver enzymes were detected frequently.36 Although we measured these very high plasma concentrations of ALLO in virtually all of our infants treated, it did not cause toxic side effects reported for this drug in the literature; no dermatologic manifestations such as pruritic, erythematous, or macopapular eruptions, or hematologic abnormalities such as leucopenia and/or eosinophilia were noticed in any patient. Although ALLO can induce elevation of transaminase activity and even hepatitis, we consider the elevation found in the infants studied to be caused by the perinatal hypoxic–ischemic insult rather than by ALLO. No differences between liver and renal function tests of treated and nontreated infants were detected, and normalization occurred equally in the surviving infants of both groups. Although several infants were treated with the combination of ALLO and amoxicillin, no dermatologic abnormalities were detected here.
Conclusions with regard to outcome cannot be drawn from this pilot study, but the (relative) preservation of CBV and ECBA in particular, during the 4- to 8-hour postnatal registration period, and the clinical course (Sarnat score) point further to less brain damage in those infants treated with ALLO.16,37 Although we are unable to confirm it because only relative changes of CBV were monitored, we suggest that the decrease in CBV represents true cerebral hypoperfusion. A growing number of studies report hypoperfusion of the brain starting 30 to 60 minutes after reperfusion after an experimentally induced hypoxic–ischemic insult.14,15,38-41 We and others found in earlier studies that a drop in brain perfusion or CBV in severely asphyxiated infants during the first 12 hours of life was related with adverse outcome.14,37 The same phenomenon has been reported extensively in newborn animals exposed to hypoxic ischemia with subsequent severe brain damage.15,38,42 Excess formation of free radicals indeed is known to induce abnormal arteriolar reactivity directly,43 which may have caused the steep decrease observed in cerebral perfusion, especially in those infants not treated with ALLO. However, evidence is growing that post hypoxic–ischemic reperfusion-induced accumulation of granulocytes, occluding a substantial percentage of the cerebral capillary bed, adds to the hypoperfusion of important regions of the brain.2,39,44Caceres et al40 showed in asphyxiated piglets that a widespread leukocyte accumulation occurred in cerebral capillaries with consequent perfusion disturbances only 4 to 6 hours after the insult, which is exactly the time period in which we detected decreases in NIRS-measured CBV in our CONT infants with adverse outcome. Moreover, a recent study by Hudome et al41 showed that neutrophils contribute to hypoxic–ischemic brain injury and that neutrophil depletion before the hypoxic–ischemic insult was neuroprotective. Because reactive oxygen species are supposed to play an important role as inflammatory mediators activating endothelium to express adhesion molecules leading to granulocyte accumulation,2,45 we speculate that ALLO treatment may, at least in part, prevent this process. This postulation is supported further by the fact that ALLO-treated infants had a better preservation of electrical brain activity and gave rise to less production of nonprotein-bound iron.
Because five of the eight infants with adverse outcome (four CONT and one ALLO-treated infant) died already before 24 hours of life or shortly thereafter, comparisons between groups of hemodynamic, functional, and chemical parameters beyond 24 hours probably are less meaningful, because the brains of the remaining infants of both groups are relatively spared. Finally, we must consider the role of phenobarbital on cerebral hemodynamics and ECBA. Although the difference was not statistically significant, substantially more infants in the CONT group (n = 7) compared with the ALLO-treated group (n = 4) received this drug during the 4- to 8-hour registration period. Phenobarbital has a negative effect on cerebral metabolic rate and causes cerebral vasoconstriction, possibly lowering CBV and ECBA, especially in the CONT group.46,47
In summary, the present study showed that prooxidant activity was lower and lipid peroxidation was reduced in plasma after ALLO treatment, although antioxidative capacity might be influenced negatively by inhibition of uric acid production. Although the dose of ALLO used in the present study caused very high concentrations in plasma, no side effects were detected. Furthermore, we found a relative preservation of postasphyxial CBV and electrical brain activity in the ALLO-treated infants. These data support the consideration of establishing a multicenter trial to investigate the value of ALLO in reducing postasphyxial brain injury.
This study was supported by the Dutch Brain Foundation. C.A.D. is a recipient of a Gisela Thier Fund grant, Leiden University, The Netherlands.
We thank Dr Jan Den Hartigh for the allopurinol concentration measurements and discussions concerning this issue, and Dr Mona Toet for her expert help in interpreting the cerebral function monitor data.
- Received February 2, 1997.
- Accepted October 10, 1997.
Reprint requests to (F.V.B.) Department of Neonatology, Wilhelmina Children's Hospital, Nieuwe Gracht 137, 3512 LK, Utrecht, The Netherlands.
- ALLO =
- allopurinol •
- CONT =
- control •
- NIRS =
- near-infrared spectroscopy •
- ECBA =
- electrocortical brain activity •
- AA/DHA =
- ascorbic acid/dehydroascorbic acid •
- TRAP =
- total peroxyl radical antioxidant trapping capacity •
- Hb =
- hemoglobin •
- ΔHbO2 =
- changes in oxygenated hemoglobin •
- ΔHbR =
- changes in deoxygenated hemoglobin •
- ΔHbtot =
- total hemoglobin •
- cerebral blood volume =
- ΔCBV, changes in cerebral blood volume •
- CPK =
- creatine phosphokinase •
- HIE =
- hypoxic–ischemic encephalopathy
- Saugstad OD,
- Aasen AO
- Palmer C,
- Smith MB,
- Williams GD,
- et al.
- Moorhouse PC,
- Grootveld M,
- Halliwell B,
- Quinlan JG,
- Gutteridge JMC
- Hurn PD,
- Koehler RC,
- Blizzard KK,
- Traystman J
- Lopez-Anaya A,
- Mayersohn M
- Halliwell B,
- Chirico S
- ↵Prior P. Monitoring cerebral function. Long-term Recordings of Cerebral Electrical Activity. Amsterdam, The Netherlands: North Holland Biomedical Press; 1979:45–301
- ↵Glantz SA, Slinker BK. Primer of Applied Regression and Analysis of Variance. New York, NY: McGraw-Hill; 1990
- Palmer C,
- Roberts RL
- Gutteridge JMC,
- Richmond B,
- Halliwell B
- Saugstad OD
- Moison RMW,
- De Beaufort AJ,
- Haasnoot AA,
- Dubbelman TMAR,
- Van Zoeren Grobben D, Berger HM
- Van Bel F,
- Dorrepaal CA,
- Benders MJNL,
- Zeeuwe PEM,
- Van De Bor M,
- Berger HM
- Leffler CW,
- Busijia DW,
- Armstead WM,
- et al.
- Dustin ML,
- Rothlein R,
- Bhan AK,
- et al.
- Copyright © 1998 American Academy of Pediatrics